I am aware of OWLIM and AllegroGraph as two triple stores that are able to generate inferred triples and put them back into the dataset. Do you know others?
Bonus question: on what subset of RDFS or OWL do they do that?
Super bonus question: what do they do with the inferred triples exactly? E.g., put everything back in the same big graph? Put the inferred triples in a named graph? ...in multiple named graphs depending on the rule triggered? ...depending on the antecedent triples used by the rule?
asked 07 Dec '12, 10:18
Antoine Zimm... ♦
Jena does, using the rule engine (in forward mode, it's RETE but it has a backward mode as well : these can be combined where forwards rules generate backwards rules).
An inference graph a union of base data and the inferred triples. This is independent of named graphs or SPARQL.
answered 07 Dec '12, 19:09
Sesame provides a forward chaining inferencer for the full RDFS closure (except, of course, the infinite set of axioms that assert
answered 10 Dec '12, 01:40
Jeen Broekstra ♦